From: Doug Gilmore <Doug.Gilmore@imgtec.com>
To: Simon Marchi <simon.marchi@polymtl.ca>,
"Maciej W. Rozycki" <macro@imgtec.com>
Cc: Luis Machado <lgustavo@codesourcery.com>, <gdb-patches@sourceware.org>
Subject: Re: [PATCH] Fix PR 21337 (v4): segfault when re-reading symbols with remote debugging.
Date: Tue, 27 Jun 2017 17:29:00 -0000 [thread overview]
Message-ID: <3abdfadd-9ec1-0f2c-45ed-de2184340320@imgtec.com> (raw)
In-Reply-To: <1314dbb1177c39791e5f7a5d51c13089@polymtl.ca>
On 06/25/2017 04:24 AM, Simon Marchi wrote:
> Hi Doug,
>
> Good thing you pinged, I had completely missed it, sorry for the wait.
>
> +gdb_test "source $gdbfile" ".*source-command-completed.*" \
> + "source $testfile.gdb"
> +# Sometimes the failure only occurs on the second invocation.
> +gdb_test "source $gdbfile" ".*source-command-completed.*" \
> + "source $testfile.gdb"
>
> Tests should have unique names, so that when one fails, you know easily which one. So, perhaps "source $testfile.gdb 1" and "source $testfile.gdb 2"?
>
> The patch is ok with this fixed.
>
> Thanks,
>
> Simon
Hi Simon and Maciej,
Simon: I updated the tests per your request. When I asked Maciej to
commit the patch for me, he noticed that the comment extended over the
soft 72 column limit of, so I reformatted it. Also I corrected a typo
and format issues in the changelog entries that he noticed.
Thanks,
Doug
Fix PR 21337: segfault when re-reading symbols.
Fix issue exposed by commit 3e29f34.
The basic issue is that section data referenced through an objfile
pointer can also be referenced via the program-space data pointer,
although via a separate mapping mechanism, which is set up by
update_section_map. Thus once section data attached to an objfile
pointer is released, the section map associated with the program-space
data pointer must be marked dirty to ensure that update_section_map is
called to prevent stale data being referenced. For the matter at hand
this marking is being done via a call to objfiles_changed.
Before commit 3e29f34 objfiles_changed could be called after all of
the objfile pointers were processed in reread_symbols since section
data references via the program-space data pointer would not occur in
the calls of read_symbols performed by reread_symbols.
With commit 3e29f34 MIPS target specific calls to find_pc_section were
added to the code for DWARF information processing, which is called
via read_symbols. Thus in reread_symbols the call to objfiles_changed
needs to be called before calling read_symbols, otherwise stale
section data can be referenced.
Thanks to Luis Machado for providing text for the main comment
associated with the change.
gdb/
2017-??-?? Doug Gilmore <Doug.Gilmore@imgtec.com>
PR gdb/21337.
* symfile.c (reread_symbols): Call objfiles_changed just before
read_symbols.
gdb/testsuite
2017-??-?? Doug Gilmore <Doug.Gilmore@imgtec.com>
PR gdb/21337
* gdb.base/reread-readsym.exp: New file.
* gdb.base/reread-readsym.c: New file.
diff --git a/gdb/symfile.c b/gdb/symfile.c
index aa53415..ce37390 100644
--- a/gdb/symfile.c
+++ b/gdb/symfile.c
@@ -2587,6 +2587,9 @@ reread_symbols (void)
/* Free the obstacks for non-reusable objfiles. */
psymbol_bcache_free (objfile->psymbol_cache);
objfile->psymbol_cache = psymbol_bcache_init ();
+
+ /* NB: after this call to obstack_free, objfiles_changed
+ will need to be called (see discussion below). */
obstack_free (&objfile->objfile_obstack, 0);
objfile->sections = NULL;
objfile->compunit_symtabs = NULL;
@@ -2639,6 +2642,23 @@ reread_symbols (void)
clear_complaints (&symfile_complaints, 1, 1);
objfile->flags &= ~OBJF_PSYMTABS_READ;
+
+ /* We are about to read new symbols and potentially also
+ DWARF information. Some targets may want to pass addresses
+ read from DWARF DIE's through an adjustment function before
+ saving them, like MIPS, which may call into
+ "find_pc_section". When called, that function will make
+ use of per-objfile program space data.
+
+ Since we discarded our section information above, we have
+ dangling pointers in the per-objfile program space data
+ structure. Force GDB to update the section mapping
+ information by letting it know the objfile has changed,
+ making the dangling pointers point to correct data
+ again. */
+
+ objfiles_changed ();
+
read_symbols (objfile, 0);
if (!objfile_has_symbols (objfile))
@@ -2671,9 +2691,6 @@ reread_symbols (void)
if (!new_objfiles.empty ())
{
- /* Notify objfiles that we've modified objfile sections. */
- objfiles_changed ();
-
clear_symtab_users (0);
/* clear_objfile_data for each objfile was called before freeing it and
diff --git a/gdb/testsuite/gdb.base/reread-readsym.c b/gdb/testsuite/gdb.base/reread-readsym.c
new file mode 100644
index 0000000..2fee696
--- /dev/null
+++ b/gdb/testsuite/gdb.base/reread-readsym.c
@@ -0,0 +1,22 @@
+/* This testcase is part of GDB, the GNU debugger.
+
+ Copyright 2017 Free Software Foundation, Inc.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 3 of the License, or
+ (at your option) any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with this program. If not, see <http://www.gnu.org/licenses/>. */
+
+int
+main (void)
+{
+ return 0;
+}
diff --git a/gdb/testsuite/gdb.base/reread-readsym.exp b/gdb/testsuite/gdb.base/reread-readsym.exp
new file mode 100644
index 0000000..b69eaad
--- /dev/null
+++ b/gdb/testsuite/gdb.base/reread-readsym.exp
@@ -0,0 +1,61 @@
+# Copyright 2017 Free Software Foundation, Inc.
+#
+# This program is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 3 of the License, or
+# (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with this program. If not, see <http://www.gnu.org/licenses/>.
+
+standard_testfile
+
+set gdbfile [standard_output_file ${testfile}.gdb]
+
+# Test rereading executable. See PR gdb/21337.
+
+proc generate_cmd_file {gdbfile binfile} {
+ set ofd [open $gdbfile w]
+
+ puts $ofd "file ${binfile}"
+ puts $ofd "shell sleep 1; touch ${binfile}"
+ puts $ofd "run"
+ puts $ofd "file"
+ puts $ofd "file ${binfile}"
+ puts $ofd "shell sleep 1; touch ${binfile}"
+ puts $ofd "run"
+ puts $ofd "file"
+ puts $ofd "file ${binfile}"
+ puts $ofd "shell sleep 1; touch ${binfile}"
+ puts $ofd "run"
+ puts $ofd "file"
+ puts $ofd "p \"source-command-completed\""
+ close $ofd
+}
+
+if [use_gdb_stub] {
+ return 0
+}
+
+if {[prepare_for_testing "failed to prepare" $testfile $srcfile debug]} {
+ return -1
+}
+
+# Start with a fresh gdb.
+clean_restart ${testfile}
+
+# Using the source command to read commands from a file is important,
+# otherwise section data is freed and reallocated using the same
+# memory locations and the bug is not exposed.
+generate_cmd_file $gdbfile $binfile
+
+gdb_test "source $gdbfile" ".*source-command-completed.*" \
+ "source $testfile.gdb 1"
+# Sometimes the failure only occurs on the second invocation.
+gdb_test "source $gdbfile" ".*source-command-completed.*" \
+ "source $testfile.gdb 2"
--
1.9.1
next prev parent reply other threads:[~2017-06-27 17:29 UTC|newest]
Thread overview: 25+ messages / expand[flat|nested] mbox.gz Atom feed top
2017-03-31 23:04 [PATCH] [mips] Fix PR 21337 v1: " Doug Gilmore
2017-04-10 16:01 ` Doug Gilmore
2017-04-12 18:52 ` Luis Machado
2017-04-12 21:42 ` Doug Gilmore
2017-04-13 18:56 ` [PATCH] Fix PR 21337 v2: " Doug Gilmore
2017-04-14 15:33 ` Luis Machado
2017-04-22 2:15 ` Simon Marchi
2017-04-25 18:16 ` Doug Gilmore
2017-04-27 19:46 ` Simon Marchi
2017-04-28 23:44 ` Doug Gilmore
2017-04-29 1:13 ` Luis Machado
2017-04-29 1:42 ` Simon Marchi
2017-04-29 17:12 ` Doug Gilmore
2017-04-29 23:26 ` Simon Marchi
2017-04-30 5:14 ` Doug Gilmore
2017-05-10 23:26 ` Doug Gilmore
2017-05-12 3:29 ` Simon Marchi
2017-05-12 19:24 ` Doug Gilmore
2017-05-16 23:11 ` [PATCH] Fix PR 21337 (v3): " Doug Gilmore
2017-06-06 16:08 ` [PING][PATCH] " Doug Gilmore
2017-06-23 18:20 ` [PING^2][PATCH] " Doug Gilmore
2017-06-25 11:25 ` [PATCH] " Simon Marchi
2017-06-27 17:29 ` Doug Gilmore [this message]
2017-06-27 21:35 ` [PATCH] Fix PR 21337 (v4): " Doug Gilmore
2017-06-28 2:01 ` Maciej W. Rozycki
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=3abdfadd-9ec1-0f2c-45ed-de2184340320@imgtec.com \
--to=doug.gilmore@imgtec.com \
--cc=gdb-patches@sourceware.org \
--cc=lgustavo@codesourcery.com \
--cc=macro@imgtec.com \
--cc=simon.marchi@polymtl.ca \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox